Mikhail Efremov created IGNITE-24640:
----------------------------------------
Summary: Fix storages stoppings on
PartitionReplicaLifecycleManager#beforeNodeStop
Key: IGNITE-24640
URL: https://issues.apache.org/jira/browse/IGNITE-24640
Project: Ignite
Issue Type: Bug
Reporter: Mikhail Efremov
Assignee: Mikhail Efremov
*Description*
After IGNITE-24384 we don't have table replicas, but table listeners and
storages in {{TableManager}}. Also after IGNITE-23741 there was removed
{{AFTER_REPLICA_STOPPED}} that leads to a situation where {{TableManager}}
stops table storages asynchronously while zone replicas still may process
requests like {{ReplicaSafeTimeSyncRequest}} from IGNITE-22620 that are being
sent frequently by a timer ~0.5s.
*Motivation*
We should fix a race between table storages and zone replicas stops.
*Definition of done*
1. {{AFTER_REPLICA_STOPPED}} is returned back;
2. {{TableManager#cleanupResources}} now checks colocation flag;
3. {{TableManager}} stops storages only after zone replica stopped on the local
event.
4. Corresponding test is written.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
